Jun Liu

Hesitant Cloud Model and Its Application in the Risk Assessment of “The Twenty-First Century Maritime Silk Road”

2016 · Mathematical Problems in Engineering

The evaluation of human environment risk is lacking quantitative data, while the qualitative knowledge cannot be easily quantified and synthesized. Furthermore, sometimes the experts are not well acknowledged with the whole indicator system or cannot …

Deep Dyna-Q: Integrating Planning for Task-Completion Dialogue Policy Learning

2018

Training a task-completion dialogue agent via reinforcement learning (RL) is costly because it requires many interactions with real users. One common alternative is to use a user simulator. However, a user simulator usually lacks the …

An Improved Strong Tracking Cubature Kalman Filter for GPS/INS Integrated Navigation Systems

2018 · Sensors

The cubature Kalman filter (CKF) is widely used in the application of GPS/INS integrated navigation systems. However, its performance may decline in accuracy and even diverge in the presence of process uncertainties. To solve the …

Optimal phase sensitivity by quantum squeezing based on a Mach–Zehnder interferometer

2019 · New Journal of Physics

Abstract A novel scheme for the enhancement of phase sensitivity based on a Mach–Zehnder interferometer (MZI) and intensity detection is proposed. With the input of bright entangled twin beams from four wave mixing (FWM), the …

Contextual Text Style Transfer

2020 · arXiv (Cornell University)

We introduce a new task, Contextual Text Style Transfer - translating a sentence into a desired style with its surrounding context taken into account. This brings two key challenges to existing style transfer approaches: ($i$) …

The Study on the Text Classification for Financial News Based on Partial Information

2020 · IEEE Access

The goal of this paper is to conduct the study on the text classification for financial news based on partial information. By a fact that an indispensable step for the efficient use of topic information …

Improvement of the DPoS Consensus Mechanism in Blockchain Based on PLTS

2021

Contemporarily, DPoS is the fastest, most effective, decentralized and flexible consensus mechanism among all consensus mechanisms. Similar to the vote of the board, the holders cast a certain number of delegates to perform verification and …

Predicting Axillary Lymph Node Metastasis in Early Breast Cancer Using Deep Learning on Primary Tumor Biopsy Slides

2021 · Frontiers in Oncology

Objectives To develop and validate a deep learning (DL)-based primary tumor biopsy signature for predicting axillary lymph node (ALN) metastasis preoperatively in early breast cancer (EBC) patients with clinically negative ALN. Methods A total of …

A Corpus-Based Sampling to Build Training Data Set for Extracting Japanese Sentence Pattern

2022

Training data set plays an important role in Natural Language Processing (NLP) or Machine Learning (ML) Tasks. In the application of NLP in Japanese language education, construction of a high-quality training data set becomes the …

Design of Hazard Management Method Based on Blockchain

2022 · 2022 IEEE 4th International Conference on Civil Aviation Safety and Information Technology (ICCASIT)

This paper proposes a blockchain-based hazard management method. Firstly, the data structure and operation types of hazards in civil aviation risk management are analyzed in detail. Secondly, a blockchain is designed to maintain all the …

BEST: A blockchain sharding scheme based on nodes performance for enhancing both security and effiency

2023 · Research Square

<title>Abstract</title> Sharding technology is seen as the most likely solution to the scalability of blockchain, by dividing nodes into smaller groups called shards that transactions can be processed in parallel. However, most sharding scheme adopted …

A White-box Implementation of SM4 with Self-equivalence Encoding

2023 · The Computer Journal

Abstract White-box implementation can ensure the security of cryptographic algorithm in white-box attack environment without changing the inputs and outputs of the original algorithm. Most existing white-box implementations construct a series of lookup tables to …

QGEval: Benchmarking Multi-dimensional Evaluation for Question Generation

2024 · arXiv (Cornell University)

Automatically generated questions often suffer from problems such as unclear expression or factual inaccuracies, requiring a reliable and comprehensive evaluation of their quality. Human evaluation is widely used in the field of question generation (QG) …

Experimental Demonstration of Quantum Nearest Centroid Classification Algorithm using Orbital Angular Momentum of Photons

2025

We experimentally demonstrate a classical simulation of the quantum nearest centroid classification algorithm using orbital angular momentum (OAM) modes of photons, achieving an impressive classification accuracy of 86.7% on the Iris dataset.

Decoupling Classifier for Boosting Few-shot Object Detection and Instance Segmentation

2025 · arXiv (Cornell University)

This paper focus on few-shot object detection~(FSOD) and instance segmentation~(FSIS), which requires a model to quickly adapt to novel classes with a few labeled instances. The existing methods severely suffer from bias classification because of …

Recovering all 119 Feynman equations on a laptop: a full score on the strict SRBench ruler at about two seconds per problem

2026 · Zenodo (CERN European Organization for Nuclear Research)

SparkOrigin recovers all 119 equations in the SRBench Feynman benchmark, scored by SRBench's own strict symbolic-solution criterion (La Cava et al., 2021) on the official public data and sampling ranges, at roughly two seconds per …

ReCoRD: Bridging the Gap between Human and Machine Commonsense Reading Comprehension

2018 · arXiv (Cornell University)

We present a large-scale dataset, ReCoRD, for machine reading comprehension requiring commonsense reasoning. Experiments on this dataset demonstrate that the performance of state-of-the-art MRC systems fall far behind human performance. ReCoRD represents a challenge for …

BERT4Rec: Sequential Recommendation with Bidirectional Encoder Representations from Transformer

2019 · arXiv (Cornell University)

Modeling users' dynamic and evolving preferences from their historical behaviors is challenging and crucial for recommendation systems. Previous methods employ sequential neural networks (e.g., Recurrent Neural Network) to encode users' historical interactions from left to …

BERT4Rec

2019

Modeling users' dynamic preferences from their historical behaviors is challenging and crucial for recommendation systems. Previous methods employ sequential neural networks to encode users' historical interactions from left to right into hidden representations for making …

Discourse-Aware Neural Extractive Text Summarization

2020

Recently BERT has been adopted for document encoding in state-of-the-art text summarization models. However, sentence-based extractive models often result in redundant or uninformative phrases in the extracted summaries. Also, long-range dependencies throughout a document are …